Quantifying the color rendition characteristics of a light source. From CRI to TM-30-15 and beyond

In school most of us have learned about glass prisms and the experiment to reveal the composition of sun light. It is the easiest way to understand the make-up of white light, a combination of lights of different wavelengths in the electromagnetic spectrum. 

The prism science experiment turned into architectural lighting across train stations in Italy.

To obtain white light from an artificial source, a combination of different emitted wavelengths is needed. This was discovered early in the 20-est century, when it also became evident that there is more than one "recipe" to produce and combine the elements in the mix, with varying results.

Such variation has led to the requirement to measure all the relevant properties of an artificial light source. One is the ability to accurately reproduce colors. Quantifying this came to be a must do when gas-discharge lamps became widely used in the 1940-1950s, with fluorescent, sodium, metal halide, high intensity discharge being the most popular.

The many flavors of artificial white light, shown as an analysis of the light spectrum. Daylight (D65) is in the top middle.

While incandescent lamps emit light by heating a wire filament to high temperature, until it glows with visible light (incandescence), gas-discharge lamps send an electrical discharge through an ionized gas. Compared to incandescent, there is a huge variety of gas-discharge lamps due to the many possible combinations of the elements, such as the type of gas (neon, argon, xenon, krypton, neon), other metals ( mercury, sodium, and metal halides) and structural parts.

Because of the wide variation in the ability to reproduce colors with the white light emitted by the many types of gas-lamps, the color rendering index (CRI) was introduced by the International Commission on Illumination (CIE) in 1974. The CRI is a one number quantification of how good (or bad) an artificial source is at reproducing colors, compared with reference standard illuminant modeled after daylight. The highest number is 100, for daylight and incandescent / halogen lamps while gas discharge lamps have between 17 and 96, with even a negative value for low-pressure sodium (the yellow kind used in street lights).

Typical CRI values of the traditional lighting technologies

The CRI number is obtained via a standard experimental and mathematical calculation procedure made on the spectrum of the analyzed light source while it illuminates a set of eight benchmark colors.

The result from illuminating these eight colors is then compared with that of the standard illuminant or daylight and from this results the CRI value of the analyzed light source, as an average of the 8 individual test scores.

A sample image from a CRI calculation software used by a family of devices that measure the properties of light.

The choice between standard illuminant or daylight results from the color temperature of the light source. For 5000K or below the standard illuminant is, used, otherwise Daylight (D65).

The CRI, with over 40 years of existence, is firmly entrenched in the lighting industry and among professionals in the filed. However, it has failed to gain much traction among the general public due to the fact that soon after its introduction it became not really relevant when making a purchase decision for lighting. The reason was the very limited number of lamp alternatives for each application, with cases when the choice was a single type of lamp. For most cases, the CRI was good enough, 80 or more, or a high CRI was not important for the application.

For example, for office or all other linear lighting almost everybody knew and used the tri-phospor linear fluorescent tubes, on the market since the 70s and with CRI over 80. For the home you had to choose from incandescent and halogen with CRI 100, for retail and other high intensity spot lighting there where the metal halide lamps with CRI min 85. Street light was reserved for high intensity and very efficient sodium lamps that although had poor CRI very few cared about it.

CRI is a standard for a quiet, settled market, where each lighting technology had its own clear place and the CRI score did not have much importance.

 By 1974, when CRI was introduced, all lighting technologies where already extensively used for their destined applications. That included the last major improvement in florescent lamps, tri-phospor, in 1970. Until late 1990s, LEDs where on the market only colored, while OLED where only in theory or prototype stage.

LED technology changed everything. For the first time, a lighting source can be used for every application and have every performance and quality level, including the ability to accurately reproduce colors.

The makeup of some fixture types for common applications, such as down-lighting, spot-lighting, office-lighting and street-lighting. All can use LEDs.

Further more, the spectrum of a LED light source can be designed with ever greater precision.

LED color spectrum can be purposely designed for an application, such as the Luxeon Crisp White Technology that delivers a natural crisp whiteness by activating Fluorescent Whitening Agents (FWAs) in paints and fabrics to attractively display merchandise in retail shops. A Crisp White LED spectrum is different than a "typical" LED but the calculated CRI is the same.

With so much complexity, the CRI with its eight colors is not comprehensive enough for an optimum quality test of so many combinations of technologies, wavelengths and color temperatures on the market today.

The CRI bechmark results in the same score not only for two LEDs with different spectrum (see Crisp White vs Typical LED above) but also for radically different spectrum of LED and Fluorescent (above and below). Applying an average from to few tested colors is the reason.

The CRI cannot also prevent designing a light source, fluorescent or LED, with the specific purpose of having a good score at those eight colors while having a poor color spectrum overall. Its extension to fourteen colors or revision from the 90s did not change this. 

It is time for a new standard that is adapted to the current and predictable future of lighting technology. As of today dozens of alternatives have been proposed but finding the new standard is not easy. There are several reasons that is holding back the CIE from announcing a replacement of CRI.

LED technology made the general public progressively more aware about technical terms. Luminous flux, color temperature, lifetime, view angle, luminous efficacy, are now on the retail packages of most lamps and in every catalog page. CRI has also gained a popularity boost, from packing, advertising and datasheets to the new trend to measure light with hand held spectrometers. 

This decision to introduce yet another wave of technical terms when the market is still adapting to the previous, big one, is one reason.

Another reason is that manufacturers and hole supply networks, with good intentions, will be put in great difficulty by a sudden change of the "rules of the game". Their products will become uncompetitive. Measurement equipment will also require a software upgrade at best or complete replacement at worst.

There are many proposals for a replacement. The most prominent is the TM-30-15, a proposal from the leading organizations in USA, the IES (Illuminating Engineering Society) and DOE (US department of Energy). Instead of Color Rendering Index, we have the Fidelity Index with the same quality scale up to 100 but with greatly increased benchmark precision due to the increase from 8 test colors to 99.

By increasing the test colors substantially only those light sources that have a high similarity with daylight or the standard illuminant, across the complete spectrum, will get a score greater than 90 or 95. In the proposal for the Fidelity Index, the IES has presented how light sources would fare with the passing to the new benchmark.

The graph shows that important number of light sources, mainly with LED and narrowband Fluorescent have a Fidelity Index below 80, while scoring at least 80 on the Color Rendering index. The level recommended as minimum required for general lighting is 80. Any such move to a to the new index should be made slowly enough to allow the industry to adapt, otherwise many products will be kicked out from the market.

To deal with those light sources that are designed to over-saturate or saturated particular colors the IES is proposing a second index, the Gamut Index to be used in combination with the Fidelity Index and equally popularized and included in datasheets and marketing materials.

A Gamut Index (Rg) of 100 means the light source does not de- or over-saturate a color pallet, for example red. For the index value from 101 to 140 a progressive over-saturation of color exists and from 100 to 60 a de-saturation. 

A colorful graph is also suggested as accompaniment to the number.

The TM-30-15 has improved benchmark precision also by featuring a gradual switch from the standard illuminant to daylight in the color temperature interval 4500K to 5500K. The direct change for one to the other was on of the causes of reduced precision of CRI.

My opinion is that the TM-30-15 has many merits but there is still work to be done to have a replacement for CRI. There should be more focus on the fact that an artificial light source has to be similar with daylight. Every difference, like over saturation of one color, should result in a penalization in the value of the replacement for the CRI and not be made into a new index that might be lost in the communication over the supply chain. Also, as a consumer, i find disturbing the act of changing my perception of products, for example food or fashion, by over and/or under saturation of colors from light sources chosen especailly for this purpose.

Further more, choosing a light source with TM-30-15 will be more confusing. For example, what too choose between a lamp with Fidelity Index (FI) 80 and Gamut Index (GI) 100 and one with FI 90 and GI 85 ? The amount of information provided by TM-30-15 is extremely useful, for a professional, but it could make the decision to choose a light source much more difficult and time consuming.

Finding a replacement for CRI is also a matter of adapting to a market where many decision makers are still unfamiliar with many of the technical terms related to lighting and find a challenge in assimilating them. There is a lot of work for experts in marketing, communication or public relationship, not only scientists...